CAELINUX as repository
- NimaM
- Topic Author
- Offline
- New Member
-
Less
More
- Posts: 1
- Thank you received: 0
15 years 8 months ago #3402
by NimaM
CAELINUX as repository was created by NimaM
Hello all,
I used CAELINUX for a while and I am very thankful of it. I would like to recommend changing it to a repository rather than just a linux distribution with CAE applications.
Some advantages are:
1- More OSes could be covered not just PCLINUX or Ubuntu and both 32-bit and 64-bit systems.
2- Many users already work with some OSes for years and it is hard for them to change it.
3- Many users just use one or few applications from current distribution.
4- Updates could be done easier. It could be done when a single application is updated and there is no reason for waiting for all of them to be updated and have a major release.
For an example of such a trend you could have a look at ccrma.stanford.edu/planetccrma/software/ . CCRMA at Stanford provides audio visual applications under linux with adjusted kernel for years. Until I guess about two years ago the user should download whole OS from their website and they were responsible for every components of it. Nowadays they only distribute audio-visual ready to install applications.
This can be the same for CAE applications. An alternative would be keeping the current form of release and adding repository support.
Best regards,
Nima
I used CAELINUX for a while and I am very thankful of it. I would like to recommend changing it to a repository rather than just a linux distribution with CAE applications.
Some advantages are:
1- More OSes could be covered not just PCLINUX or Ubuntu and both 32-bit and 64-bit systems.
2- Many users already work with some OSes for years and it is hard for them to change it.
3- Many users just use one or few applications from current distribution.
4- Updates could be done easier. It could be done when a single application is updated and there is no reason for waiting for all of them to be updated and have a major release.
For an example of such a trend you could have a look at ccrma.stanford.edu/planetccrma/software/ . CCRMA at Stanford provides audio visual applications under linux with adjusted kernel for years. Until I guess about two years ago the user should download whole OS from their website and they were responsible for every components of it. Nowadays they only distribute audio-visual ready to install applications.
This can be the same for CAE applications. An alternative would be keeping the current form of release and adding repository support.
Best regards,
Nima
- Joël Cugnoni
-
- Offline
- Moderator
-
15 years 7 months ago #3433
by Joël Cugnoni
Joël Cugnoni - a.k.a admin
www.caelinux.com
Replied by Joël Cugnoni on topic Re:CAELINUX as repository
Yes i agree with you:
a repository would complement well the current Live DVD approach. But unfortunatelly, packaging these simulation codes is a big task, much more complicated than packaging most other programs.
Even so, providing binary packages is possible, but ensuring that it works on a wide range of distribution is even more complicated.
There are two options then:
1) restrict the repository to certain distro (Ubuntu/Debian?)
2) develop "stand alone" binary packages containing ALL required libraries (=huge)
For example, EDF has packaged Salome-MECA & Aster in the form of a "standalone" binary archive, and it works well on most distro, but it weights approx 600 Mb and cannot be updated through your prefered package manager. The same for OpenFOAM: works well, but it's huge!
On the other hand, there exist Debian and/or RPM packages for Elmer, Gerris. Thes packages are very convenient to install but they are specific to one type of distribution.
In the future, I will try to develop DEB packages of the codes that I include in CAELinux. It will at least allow users of Ubuntu (or Debian) to install and upgrade easily. However, some codes require specific configuration by the user and this can be hardly automated (amount of RAM, location of scratch files, MPI & SSH options). Pre-compiled is on thing, Pre-configured and ready to run is another thing. That's also why I want to keep the LiveDVD system: everything can be pre-configured to work "out of the box"!
In the meantime, I strongly encourage people who know how to package codes to try to develop Ubuntu packages of the codes they are interested in!
a repository would complement well the current Live DVD approach. But unfortunatelly, packaging these simulation codes is a big task, much more complicated than packaging most other programs.
Even so, providing binary packages is possible, but ensuring that it works on a wide range of distribution is even more complicated.
There are two options then:
1) restrict the repository to certain distro (Ubuntu/Debian?)
2) develop "stand alone" binary packages containing ALL required libraries (=huge)
For example, EDF has packaged Salome-MECA & Aster in the form of a "standalone" binary archive, and it works well on most distro, but it weights approx 600 Mb and cannot be updated through your prefered package manager. The same for OpenFOAM: works well, but it's huge!
On the other hand, there exist Debian and/or RPM packages for Elmer, Gerris. Thes packages are very convenient to install but they are specific to one type of distribution.
In the future, I will try to develop DEB packages of the codes that I include in CAELinux. It will at least allow users of Ubuntu (or Debian) to install and upgrade easily. However, some codes require specific configuration by the user and this can be hardly automated (amount of RAM, location of scratch files, MPI & SSH options). Pre-compiled is on thing, Pre-configured and ready to run is another thing. That's also why I want to keep the LiveDVD system: everything can be pre-configured to work "out of the box"!
In the meantime, I strongly encourage people who know how to package codes to try to develop Ubuntu packages of the codes they are interested in!
Joël Cugnoni - a.k.a admin
www.caelinux.com
- dust
- Offline
- New Member
-
Less
More
- Posts: 7
- Thank you received: 0
15 years 7 months ago #3440
by dust
Replied by dust on topic Re:CAELINUX as repository
launchpad.net/~cae-team
i think we can get more space
answers.launchpad.net/soyuz/+question/74277
how much space do you think we would need?
launchpad.net/~cae-team/+archive/ppa
that is what is there now
my suggestion is todo it similar to ubuntustudio, a metapackage with all included
launchpad.net/~ubuntustudio-dev/+related-software
i think we can get more space
answers.launchpad.net/soyuz/+question/74277
how much space do you think we would need?
launchpad.net/~cae-team/+archive/ppa
that is what is there now
my suggestion is todo it similar to ubuntustudio, a metapackage with all included
launchpad.net/~ubuntustudio-dev/+related-software
- dust
- Offline
- New Member
-
Less
More
- Posts: 7
- Thank you received: 0
15 years 5 months ago #3620
by dust
Replied by dust on topic Re:CAELINUX as repository
now that ubuntu 9.10 is released we would need help at launchpad.net/~cae-team/+archive/ppa to bring complete caelinux to ubuntu users.
- Brenda EM
- Offline
- New Member
-
Less
More
- Posts: 18
- Thank you received: 0
15 years 5 months ago #3663
by Brenda EM
Replied by Brenda EM on topic Re:CAELINUX as repository
I too hope that a Salome/Aster/Code_Saturne repository is created. Other people are also working on this too:
www.salome-platform.org/forum/forum_9/504522495
I understand the this might be a provocative statement, but: If a piece of software is so difficult to install that it cannot be maintained by a normal package management scheme, perhaps instead, it should be fixed or changed to do so.
What would be ideal is for Salome and Gmsh to be interchangeable, upgradability, and modularity, to share the same binary versions of Netgen and Tetgen, for instance, if that condition does not exist yet.
I wish also that Salome did not need to be compiled to encompass Aster and Saturne, but rather check for the presence of them, and use them if they are there.
www.salome-platform.org/forum/forum_9/504522495
I understand the this might be a provocative statement, but: If a piece of software is so difficult to install that it cannot be maintained by a normal package management scheme, perhaps instead, it should be fixed or changed to do so.
What would be ideal is for Salome and Gmsh to be interchangeable, upgradability, and modularity, to share the same binary versions of Netgen and Tetgen, for instance, if that condition does not exist yet.
I wish also that Salome did not need to be compiled to encompass Aster and Saturne, but rather check for the presence of them, and use them if they are there.
- Jake
- Offline
- Senior Member
-
Less
More
- Posts: 49
- Thank you received: 0
15 years 5 months ago #3667
by Jake
Replied by Jake on topic Re:CAELINUX as repository
May I ask why gmsh? I have found SALOME to be more reliable when it comes to meshing and a lot more stable when it comes to the actual meshing processing part. Gmsh has always locked up on me and therefore been useless.
In response to the original post, I really support the idea of both 32-bit and 64-bit versions. I am a student and really don't have the money to buy another computer, let alone a high-end one. I was very disappointed when there was a 64-bit only version for CAELinux 2009. I didn't want to use CAELinux 2008 because it's all outdated now and ndiswrapper doesn't support my wireless card.
In response to the original post, I really support the idea of both 32-bit and 64-bit versions. I am a student and really don't have the money to buy another computer, let alone a high-end one. I was very disappointed when there was a 64-bit only version for CAELinux 2009. I didn't want to use CAELinux 2008 because it's all outdated now and ndiswrapper doesn't support my wireless card.
Moderators: catux
Time to create page: 0.125 seconds